jpmschweitzer/settled-reach
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(simulation+ui): Atlas depth — regional-map design (D-226), climate overlay, quarter glyphs #185

Closed
jpmschweitzer wants to merge 0 commits from atlas-depth into main
pull from: atlas-depth
merge into: :main
Conversation 1 Commits - Files Changed -
+1784 -321
jpmschweitzer commented 2026-07-18 12:30:00 +02:00
Owner
Copy Link
Copy Source

Summary

The Atlas depth batch (D-226): the regional-map design record plus two new viewer overlays. After this merges, the Atlas shows climate (mean-temp) at region granularity and settlement quarter-footprint glyphs, and the jet-plane-altitude regional view is fully designed with its wiring ticketed (T-1137/T-1138).

T-1124 — D-226 amendment: windowed district-resolution regional map (design, Tyre + Araminta)

Five sections recorded and validated:

  1. Request: window_center/window_n as serde-default fields ON AtlasLayerRequest — absent = whole-body; a new inbound wire shape is forbidden (PR #184 five-shape ceiling).
  2. Response carrier (ruling): district_window: Option<DistrictWindowLayer> is a windowed viewport query keyed (body, center, n) with an echoed-center staleness guard — deliberately OUTSIDE the dense whole-body layer family; that family's growth ceiling is re-scoped (not busted) and still stands.
  3. Fields: six per-cell integers — morphology, elev_q, temp_dc (i16, region-sentinel), moisture_q, vegetation (Marine-exhaustive per T-1126), glaciation_grade (T-1127's deferred wire half, now accepted).
  4. Budget (binding): 7 bytes/cell; MAX_N=64 (~28 KiB, matches the T-1123 proven window), DEFAULT_N=32; pan-only re-request (zoom never refetches); client LRU cache valid forever per D-227 determinism; temp shares ONE colorizer with the region overlay across zoom levels (D-243 continuity).
  5. Screen (Araminta): zoom-threshold LOD swap in the existing AtlasViewer (no new screen); morphology base lightness-modulated by elev_q; gen_dw_* toggles; glaciation as always-on ice-tint modifier; Marine transparent; implant chrome discipline (map palettes stay out of ACCENT_ACTIVE gold).

Follow-ups filed: T-1137 (server wiring — carries the named TerrainAnalysis cache-vs-rederive decision: derive_district needs the ~2 MB transient analysis BodyWorldState deliberately drops), T-1138 (client screen, incl. the zoom-headroom note).

T-1119 — QuarterFootprintLayer (server: Dudley, client: Stig)

  • ZoningType gains Ord (T-994 mirror); build_quarter_footprint_layer — per-city aggregates (density floor-mean, dominant type/zoning by mode with declaration-order tie-break, landmark/corridor counts), None on empty quarters, async-lag placements SKIPPED not defaulted; sixth and final whole-body response field, growth-ceiling doc updated.
  • Tie-break bug caught by its own test: Iterator::max_by_key returns the LAST max on ties over a BTreeMap iterator — backwards from declaration-order-first; replaced with an explicit forward walk.
  • Client: gen_l4_quarters (QTR) — density-scaled glyph on the L3 settlement dot, shape = dominant district type, settlement-gold density ramp, zoom-gated; landmark/corridor never drawn (D-226(d) tooltip-only ceiling).
  • Known-inert field, documented: landmark_count computes 0 on real data until landmark placement lands (BlockSkeleton.landmark is hardcoded None today); fixtures exercise the logic with non-zero counts.

T-1118 — climate overlay (client, Stig)

gen_region_grid (TMP): mean-temp cold→hot ramp over −50..+50 °C at region granularity; the airless sentinel (REGION_TEMP_NONE_DC = i16::MIN) is skip-cell — undrawn, never an invented color. Populated RegionGridLayer wire fixture added (was nil in all three fixtures since PR #179) and the Tier-2 replay test asserts exact literals from the real server-generated fixture, including the sentinel.

Structural cleanups riding along

  • atlas_viewer.gd and protocol.gd were both over gdlint's 1000-line cap pre-batch; cleanly-separable responsibilities extracted on existing precedent: atlas_generation_state.gd, atlas_generation_proxy.gd, atlas_overlay_colors.gd, atlas_map_protocol.gd (the browse_protocol delegate pattern). Public APIs preserved exactly; both files now under cap.
  • Stale TODO(#957) removed (premise predated T-957 landing).
  • T-1139 filed (pre-existing msgpack Vec-as-generic-Array decode gap, found while wiring).

Verification

  • Full client suite 3094/3094 (incl. new pure-function suites for ramp + glyph and exact-literal replay); server atlas:: 549 green; both via the pre-push gate on push.
  • Live capture (tests/run-visual) against a running server: legend grows to 7 sections, TMP/QTR toggle cleanly, no errors; glyph pixel-truth is pinned by the exact-value fixture test (the capture body's async region/quarter stages hadn't populated within the tick budget — the fixture test is the stronger check).
  • pql decisions validate clean on the amendment.

Tickets: T-1124, T-1118, T-1119

## Summary The Atlas depth batch (D-226): the regional-map design record plus two new viewer overlays. After this merges, the Atlas shows climate (mean-temp) at region granularity and settlement quarter-footprint glyphs, and the jet-plane-altitude regional view is fully designed with its wiring ticketed (T-1137/T-1138). ### T-1124 — D-226 amendment: windowed district-resolution regional map (design, Tyre + Araminta) Five sections recorded and validated: 1. **Request**: `window_center`/`window_n` as serde-default fields ON `AtlasLayerRequest` — absent = whole-body; a new inbound wire shape is forbidden (PR #184 five-shape ceiling). 2. **Response carrier (ruling)**: `district_window: Option<DistrictWindowLayer>` is a *windowed viewport query* keyed `(body, center, n)` with an echoed-center staleness guard — deliberately OUTSIDE the dense whole-body layer family; that family's growth ceiling is re-scoped (not busted) and still stands. 3. **Fields**: six per-cell integers — morphology, elev_q, temp_dc (i16, region-sentinel), moisture_q, vegetation (Marine-exhaustive per T-1126), glaciation_grade (T-1127's deferred wire half, now accepted). 4. **Budget (binding)**: 7 bytes/cell; MAX_N=64 (~28 KiB, matches the T-1123 proven window), DEFAULT_N=32; pan-only re-request (zoom never refetches); client LRU cache valid forever per D-227 determinism; temp shares ONE colorizer with the region overlay across zoom levels (D-243 continuity). 5. **Screen (Araminta)**: zoom-threshold LOD swap in the existing AtlasViewer (no new screen); morphology base lightness-modulated by elev_q; `gen_dw_*` toggles; glaciation as always-on ice-tint modifier; Marine transparent; implant chrome discipline (map palettes stay out of ACCENT_ACTIVE gold). Follow-ups filed: T-1137 (server wiring — carries the named TerrainAnalysis cache-vs-rederive decision: `derive_district` needs the ~2 MB transient analysis BodyWorldState deliberately drops), T-1138 (client screen, incl. the zoom-headroom note). ### T-1119 — QuarterFootprintLayer (server: Dudley, client: Stig) - `ZoningType` gains Ord (T-994 mirror); `build_quarter_footprint_layer` — per-city aggregates (density floor-mean, dominant type/zoning by mode with declaration-order tie-break, landmark/corridor counts), None on empty quarters, async-lag placements SKIPPED not defaulted; sixth and final whole-body response field, growth-ceiling doc updated. - **Tie-break bug caught by its own test**: `Iterator::max_by_key` returns the LAST max on ties over a BTreeMap iterator — backwards from declaration-order-first; replaced with an explicit forward walk. - Client: `gen_l4_quarters` (QTR) — density-scaled glyph on the L3 settlement dot, shape = dominant district type, settlement-gold density ramp, zoom-gated; landmark/corridor never drawn (D-226(d) tooltip-only ceiling). - Known-inert field, documented: `landmark_count` computes 0 on real data until landmark placement lands (BlockSkeleton.landmark is hardcoded None today); fixtures exercise the logic with non-zero counts. ### T-1118 — climate overlay (client, Stig) `gen_region_grid` (TMP): mean-temp cold→hot ramp over −50..+50 °C at region granularity; the airless sentinel (`REGION_TEMP_NONE_DC` = i16::MIN) is skip-cell — undrawn, never an invented color. Populated RegionGridLayer wire fixture added (was nil in all three fixtures since PR #179) and the Tier-2 replay test asserts exact literals from the real server-generated fixture, including the sentinel. ### Structural cleanups riding along - `atlas_viewer.gd` and `protocol.gd` were both over gdlint's 1000-line cap pre-batch; cleanly-separable responsibilities extracted on existing precedent: `atlas_generation_state.gd`, `atlas_generation_proxy.gd`, `atlas_overlay_colors.gd`, `atlas_map_protocol.gd` (the browse_protocol delegate pattern). Public APIs preserved exactly; both files now under cap. - Stale `TODO(#957)` removed (premise predated T-957 landing). - T-1139 filed (pre-existing msgpack Vec<u8>-as-generic-Array decode gap, found while wiring). ### Verification - Full client suite 3094/3094 (incl. new pure-function suites for ramp + glyph and exact-literal replay); server `atlas::` 549 green; both via the pre-push gate on push. - Live capture (`tests/run-visual`) against a running server: legend grows to 7 sections, TMP/QTR toggle cleanly, no errors; glyph pixel-truth is pinned by the exact-value fixture test (the capture body's async region/quarter stages hadn't populated within the tick budget — the fixture test is the stronger check). - `pql decisions validate` clean on the amendment. Tickets: T-1124, T-1118, T-1119
jpmschweitzer added 3 commits 2026-07-18 12:30:00 +02:00
docs(meta): D-226 T-1124 amendment — windowed district-resolution regional map design (T-1124) 67bd0ec294 
Five sections, Tyre data shape + Araminta screen/encoding (T-1112
pattern): (1) request = serde-default window_center/window_n ON
AtlasLayerRequest (five-shape demux cap honored); (2) response carrier
RULING — district_window is a windowed viewport query keyed
(body, center, n) with echoed-center staleness guard, deliberately
OUTSIDE the dense whole-body layer family whose growth ceiling is
re-scoped (not busted) to that family; (3) six per-cell fields incl.
glaciation_grade (T-1127 deferral accepted) and Marine-exhaustive
vegetation; (4) binding budget — 7 bytes/cell, MAX_N=64 (~28 KiB),
DEFAULT_N=32, temp shares the T-1118 region-ramp colorizer across zoom
levels (D-243 continuity), pan-only re-request (zoom never refetches),
D-227-determinism client cache; (5) screen — zoom-threshold LOD swap
in the existing AtlasViewer, morphology base with elev_q lightness,
gen_dw_* toggles, glaciation as always-on tint modifier, Marine
transparent, implant chrome discipline.

Validated via pql decisions validate. Implementation is follow-up
ticket scope (server wiring must resolve the named TerrainAnalysis
transience decision; client screen).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
feat(simulation): T-1119 QuarterFootprintLayer — server half (touch points 1-2) + populated wire fixtures 8165c61186 
ZoningType gains PartialOrd/Ord (DistrictType T-994 mirror) for the
dominant_zoning declaration-order tie-break. layer_proxy:
QuarterFootprintEntry/QuarterFootprintLayer (BTreeMap keyed by
city_id; density_avg_pct floor-mean, dominant_* mode with
declaration-order tie-break, landmark/corridor counts) +
build_quarter_footprint_layer gated None on empty state.quarters;
sixth AtlasLayerResponse field wired into all construction sites;
growth-ceiling doc updated to say the whole-body budget is consumed
(what carries a seventh is T-1124's design, resolved in the same
branch's amendment). Placements whose derived Layer4Quarter id is not
yet in state.quarters are SKIPPED (skeleton gen is async Low-priority
— entries may be a proper subset; client decodes defensively). Stale
TODO(#957) removed (premise predates T-957 landing).

Tie-break bug caught by its own test: Iterator::max_by_key over a
BTreeMap returns the LAST max on ties — backwards from
declaration-order-first; replaced with an explicit forward walk that
replaces only on strictly-greater count.

gen_fixtures: populated RegionGridLayer (T-1118 replay coverage) +
QuarterFootprintLayer literals in the ready fixture; msgpack fixtures
regenerated via make fixtures. Note: landmark_count is inert on real
data today (BlockSkeleton.landmark hardcoded None until landmark
placement lands); fixtures use non-zero counts to exercise the logic.

5 new unit tests + populated round-trip; atlas:: 549 green.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
feat(ui): T-1118 climate overlay + T-1119 quarter glyphs — client halves, with file-size extractions e41cca2ded 
T-1118: gen_region_grid overlay (label TMP) — _draw_gen_region_grid
copies _draw_gen_district's self-contained mapping (dims from the
layer dict); mean-temp cold-to-hot ramp over -50..+50 C;
REGION_TEMP_NONE_DC airless sentinel = skip-cell (undrawn, never an
invented color); legend entry.

T-1119 (touch points 3-6): quarter_footprints protocol passthrough;
gen_l4_quarters overlay (label QTR) — density-scaled glyph anchored
on the L3 settlement dot joined by city_id, shape = dominant district
type (corner-tab/diamond marks for Commercial/Industrial/
Administrative), color = density ramp on the settlement-gold family,
zoom-gated at SETTLEMENT_LABEL_MIN_ZOOM; landmark/corridor counts
never drawn (D-226(d) tooltip-only ceiling); legend entry.

Structure: atlas_viewer.gd and protocol.gd were over gdlint's
1000-line cap before this batch; cleanly-separable responsibilities
extracted on existing precedent — atlas_generation_state.gd (per-layer
data + accessors), atlas_generation_proxy.gd (polling/retry/pending
machinery), atlas_overlay_colors.gd (pure ramp/shape lookups),
atlas_map_protocol.gd (atlas/starmap/citynames codec, the
browse_protocol.gd delegate pattern). Public APIs preserved exactly;
_gen_state stays a field default (RefCounted, pre-_ready safe) because
_ready()-construction breaks every bare AtlasViewer.new() test —
documented inline.

Tests: registration + round-trip for both overlays; pure-function
suites for the temp ramp (endpoints/midpoint/clamp/sentinel) and
quarter glyph (scaling, zoom gate, ramp, notch across all 9
DistrictTypes); Tier-2 replay asserts exact literals from the real
server-generated fixture incl. the airless sentinel. Color.lerp(a,b,
1.0) is not bit-exact to b — endpoint assertions use per-component
is_equal_approx. Full suite 3094/3094; gdlint zero warnings incl. the
two previously-over-cap files. Live capture: legend grows to 7
sections, TMP/QTR toggles clean against a live server.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
jpmschweitzer added 1 commit 2026-07-18 12:45:39 +02:00
docs(meta): D-226 T-1124 amendment — PR #185 Troblum findings (windowed-family ceiling, queue-based serving, comment reconciliation) c20c175af1 
All five findings applied, none touching the option-(c) carrier choice:
(1) SS2 gains a binding windowed-family ceiling — exactly ONE
windowed-query field; a second windowed query is a dedicated response
message by rule (two concurrent windowed payloads would need the
per-field request-correlation that IS the tagged framing — that is the
trigger to build it); symmetric with the request side's five-hard cap.
(2) SS1 gains a binding serving model — window derives ride the same
Rayon queue + Pending pattern as whole-body cache misses, never inline
in the PreInput drain (which the codebase deliberately keeps expensive
work off); per-connection coalescing recommended; SS4's 7/29ms figures
reframed as worker latency, not tick-thread cost. (3) The
layer_proxy.rs growth-ceiling comment (comment-only diff) now records
the T-1124 ruling as RESOLVED — scoped to the dense whole-body family,
district_window outside it, second windowed field = dedicated-message
migration — instead of deferring to an open design question. (4) SS2
documents that body scoping rides the enclosing AtlasLayerResponse
envelope (cache key body_id+center+n; echo disambiguates within-body).
(5) SS5 entry clarified: zoom threshold alone triggers the swap
(works over open ocean), first window centers on the pan-center's
derived DistrictPos; downstream settlement-anchoring phrasing
reconciled.

pql decisions validate clean.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
jpmschweitzer commented 2026-07-18 12:46:41 +02:00
Author
Owner
Copy Link
Copy Source

Review: atlas-depth → main (type: code + governance) — round 2 (fix round)

Three reviewers: Tyre authored the D-226 T-1124 amendment in this diff, so he reviewed only the implementation; Troblum ran the adversarial pass on the amendment.

Hoshe (QA): APPROVE

Independently verified every claim: the mode_by_declaration_order tie-break fix is a correct forward-walk (confirmed via standalone compiled trace); fixture literals match byte-for-byte what the Tier-2 replay test asserts against the committed msgpack (44/44); the four extractions faithfully copy their precedents with no autoload class_name hazards and no broken external callers; both over-cap files genuinely under cap (1017→985, 996→916); renderers use the correct self-contained mapping and never draw landmark/corridor (D-226(d)); the legend gate test is real and enforced; the TODO(#957) removal verified as stale-comment hygiene, not regression cover. Non-blocking process note: branch changelog predates T-1137/38/39 on main — resolved by the merge keeping main's copy (branch never touched changelog files); store verified post-merge.

# File Issue
— — none

Tyre (Architecture, implementation only): APPROVE

The four extraction seams are clean and match their cited precedents; decode semantics traced byte-identical end to end. QuarterFootprintLayer BTreeMap keyed city_id (D-010), city_id→quarter_id re-derivation matches the dispatch path exactly, async-subset skip sound. Enum-serialization split between sibling layers cleared as principled idiom (dense positional arrays pack as-u8 for size; sparse keyed records self-describe as named variants). T-1118's ramp lives reusably in atlas_overlay_colors.region_temp_color() — the single colorizer the future district window requires per amendment §4.

# File Issue
— — none

Troblum (adversarial, amendment): APPROVE (round 1: REQUEST_CHANGES, 3 blocking + 2 doc findings — all fixed and confirmed)

Round-1 findings, all applied in c20c175af and re-verified against the diff + underlying code:

  1. Windowed-family ceiling missing → §2 now pins exactly ONE windowed-query field; a second windowed query is a dedicated response message by rule, with the trigger grounded (concurrent windowed payloads need per-field correlation, which IS the tagged framing). Symmetric with the request side's five-hard cap.
  2. Inline derive on the tick thread (highest severity) → binding serving model: window derives ride the same Rayon queue + Pending pattern as whole-body cache misses, never inline in the PreInput drain; per-connection coalescing as the server-side backstop; 7/29ms reframed as worker latency. Verified the cited drain-doc text is genuine pre-existing code, not fabricated authority.
  3. layer_proxy.rs ceiling comment contradicted the amendment → rewritten (comment-only, struct untouched): cap scoped to the dense whole-body family, T-1124 recorded as RESOLVED, windowed cap stated inline.
  4. Staleness echo → body scoping documented as riding the response envelope (verified against actual AtlasResponseBuffer routing); echo narrowed to within-body disambiguation.
  5. First-entry/centering → zoom-threshold-alone trigger (works over open ocean), first window centers on the pan-center's derived DistrictPos, downstream phrasing reconciled.

The option-(c) carrier choice itself survived the adversarial pass unchanged ("well-reasoned… real, not a rationalization").

# Section Issue Status
1 Amendment §2 Windowed family had no cap — governance gap FIXED (c20c175af)
2 Amendment §1/§4 Inline derive in PreInput drain — tick-stall hazard FIXED — queue-based serving, binding
3 layer_proxy.rs:77-83 Code comment contradicted the ruling FIXED — reconciled, comment-only
4 Amendment §2 Staleness echo under-documented on body scoping FIXED
5 Amendment §5 First-fetch centering/trigger under-specified FIXED

Verdict: APPROVED

Runtime verification: full cargo + gdUnit suites via the pre-push gate on every push (3094 client / full server). Live capture against a running server: legend grows to 7 sections, TMP/QTR toggle clean; glyph truth pinned by exact-literal fixture replay. pql decisions validate clean on the final amendment. T-1137 updated with the binding queue-serving model before any wiring work can pick it up.

## Review: atlas-depth → main (type: code + governance) — round 2 (fix round) Three reviewers: Tyre authored the D-226 T-1124 amendment in this diff, so he reviewed only the implementation; Troblum ran the adversarial pass on the amendment. ### Hoshe (QA): APPROVE Independently verified every claim: the `mode_by_declaration_order` tie-break fix is a correct forward-walk (confirmed via standalone compiled trace); fixture literals match byte-for-byte what the Tier-2 replay test asserts against the committed msgpack (44/44); the four extractions faithfully copy their precedents with no autoload class_name hazards and no broken external callers; both over-cap files genuinely under cap (1017→985, 996→916); renderers use the correct self-contained mapping and never draw landmark/corridor (D-226(d)); the legend gate test is real and enforced; the TODO(#957) removal verified as stale-comment hygiene, not regression cover. Non-blocking process note: branch changelog predates T-1137/38/39 on main — resolved by the merge keeping main's copy (branch never touched changelog files); store verified post-merge. | # | File | Issue | |---|------|-------| | — | — | none | ### Tyre (Architecture, implementation only): APPROVE The four extraction seams are clean and match their cited precedents; decode semantics traced byte-identical end to end. QuarterFootprintLayer BTreeMap keyed city_id (D-010), city_id→quarter_id re-derivation matches the dispatch path exactly, async-subset skip sound. Enum-serialization split between sibling layers cleared as principled idiom (dense positional arrays pack as-u8 for size; sparse keyed records self-describe as named variants). T-1118's ramp lives reusably in atlas_overlay_colors.region_temp_color() — the single colorizer the future district window requires per amendment §4. | # | File | Issue | |---|------|-------| | — | — | none | ### Troblum (adversarial, amendment): APPROVE (round 1: REQUEST_CHANGES, 3 blocking + 2 doc findings — all fixed and confirmed) Round-1 findings, all applied in c20c175af and re-verified against the diff + underlying code: 1. **Windowed-family ceiling missing** → §2 now pins exactly ONE windowed-query field; a second windowed query is a dedicated response message by rule, with the trigger grounded (concurrent windowed payloads need per-field correlation, which IS the tagged framing). Symmetric with the request side's five-hard cap. 2. **Inline derive on the tick thread** (highest severity) → binding serving model: window derives ride the same Rayon queue + Pending pattern as whole-body cache misses, never inline in the PreInput drain; per-connection coalescing as the server-side backstop; 7/29ms reframed as worker latency. Verified the cited drain-doc text is genuine pre-existing code, not fabricated authority. 3. **layer_proxy.rs ceiling comment contradicted the amendment** → rewritten (comment-only, struct untouched): cap scoped to the dense whole-body family, T-1124 recorded as RESOLVED, windowed cap stated inline. 4. Staleness echo → body scoping documented as riding the response envelope (verified against actual AtlasResponseBuffer routing); echo narrowed to within-body disambiguation. 5. First-entry/centering → zoom-threshold-alone trigger (works over open ocean), first window centers on the pan-center's derived DistrictPos, downstream phrasing reconciled. The option-(c) carrier choice itself survived the adversarial pass unchanged ("well-reasoned… real, not a rationalization"). | # | Section | Issue | Status | |---|---------|-------|--------| | 1 | Amendment §2 | Windowed family had no cap — governance gap | FIXED (c20c175af) | | 2 | Amendment §1/§4 | Inline derive in PreInput drain — tick-stall hazard | FIXED — queue-based serving, binding | | 3 | layer_proxy.rs:77-83 | Code comment contradicted the ruling | FIXED — reconciled, comment-only | | 4 | Amendment §2 | Staleness echo under-documented on body scoping | FIXED | | 5 | Amendment §5 | First-fetch centering/trigger under-specified | FIXED | ### Verdict: APPROVED Runtime verification: full cargo + gdUnit suites via the pre-push gate on every push (3094 client / full server). Live capture against a running server: legend grows to 7 sections, TMP/QTR toggle clean; glyph truth pinned by exact-literal fixture replay. pql decisions validate clean on the final amendment. T-1137 updated with the binding queue-serving model before any wiring work can pick it up.
jpmschweitzer closed this pull request 2026-07-18 12:47:16 +02:00

Pull request closed

This pull request cannot be reopened because the branch was deleted.
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
No items
No labels
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Assignees
Clear assignees
jpmschweitzer (Jeroen Schweitzer)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: jpmschweitzer/settled-reach#185